For humans, vision is the most direct and effective approach for mankind to perceive external environment. The purpose of studying computer vision is to expect that computers can perceive various states in the environment like human vision, such as color, shape, surface textual information of an object, environmental position of the object, motion condition of the object, and the like. Depth perception technique, as a hot issue in computer vision study and application development, plays a crucial role in fields like virtual reality, three-dimensional rebuilding, human-machine interaction, industrial automation, robot navigation, and medical imaging. For example, in the field of consumer electronics, the depth perception technology can help an electronic product to recognize different actions of human beings to thereby make corresponding reactions, which brings a brand-new experience that is more convenient and intelligent to the user and thus plays a significant role in enhancing the interaction competence and intelligent level of the electronic product. In the industrial field, sectors such as 3D printing, robot manipulation, also have urgent needs of applying high-precision and high-resolution depth information technologies.
Depth perception technologies may be divided into active types and passive types based on whether an active light source exists in a depth perception device. An active-type depth perception technology uses an image sensor to capture a pattern projected by a projector onto a spatial object; the pattern is analyzed and processed to obtain spatial position information of the object, e.g., structured light encoded manner, ToF flight time manner. The active-type depth perception technology are featured in that the depth information is stable, reliable, and less affected by ambient light, and a matching process is not affected by textual information of the object. For a structured light encoding-based depth perception method, the design of an encoded pattern projector is the essential part, and the designed encoded pattern will directly affect the computational complexity of depth decoding, as well as the precision and resolution of the depth information.